Scholarlink

<p>The increasing growth of Internet of Things (IoT) devices has escalated the surge for efficient task offloading techniques. Traditional bounded tasks offloading techniques encounter problems associated with dynamic workloads, low latency and using multiple memory locations inefficiently. The paper proposes a new task offloading principles in two phase for Edge-IoT systems which uses Adversarial Neuromorphic Orchestration (ANO) for task identification and routing, and Holographic Memory-based Offloading (HMO) task reuse in addition to memory usage efficiency. The proposed method allows for dynamic actions to be made based on real time resource conditions while using biological principles of neural and memory systems for intelligent scheduling for edge-offloading. The results show that ANO + HMO reduced average latency by 24.7%, reduced average energy usage by 14.9%, with an average drop rate of 41.5% compared to current methods such as Deep Q-Network (DQN) and Multi-Access Edge Computing—Particle Swarm Optimization (MEC-PSO). The proposed approach also led to 84.7% memory reuse efficiency, the best compared with existing baselines in the field. In addition, the proposed method demonstrated 95.4% task success rate, and our model can scale robustly and reliably for different IoT workloads. Overall, ANO + HMO(TPOTO) is a versatile biological systems-based solution for real-time adaptive task offloading in heterogeneous edge computing systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Two-phase real-time task offloading framework for edge-IoT systems using spiking neuromorphic coordination and holographic memory reuse

  • 期刊论文

  • 发表日期：

    2025年9月10日

查看全文
  • Paritosh Kumar Yadav,
  • Sudhakar Pandey,
  • Vikash Kumar

摘要

The increasing growth of Internet of Things (IoT) devices has escalated the surge for efficient task offloading techniques. Traditional bounded tasks offloading techniques encounter problems associated with dynamic workloads, low latency and using multiple memory locations inefficiently. The paper proposes a new task offloading principles in two phase for Edge-IoT systems which uses Adversarial Neuromorphic Orchestration (ANO) for task identification and routing, and Holographic Memory-based Offloading (HMO) task reuse in addition to memory usage efficiency. The proposed method allows for dynamic actions to be made based on real time resource conditions while using biological principles of neural and memory systems for intelligent scheduling for edge-offloading. The results show that ANO + HMO reduced average latency by 24.7%, reduced average energy usage by 14.9%, with an average drop rate of 41.5% compared to current methods such as Deep Q-Network (DQN) and Multi-Access Edge Computing—Particle Swarm Optimization (MEC-PSO). The proposed approach also led to 84.7% memory reuse efficiency, the best compared with existing baselines in the field. In addition, the proposed method demonstrated 95.4% task success rate, and our model can scale robustly and reliably for different IoT workloads. Overall, ANO + HMO(TPOTO) is a versatile biological systems-based solution for real-time adaptive task offloading in heterogeneous edge computing systems.